I am looking for a dataset mapping concrete words with common colors. The only one i've found is this, but it has less than 100 entries.

EDIT: example:

red_words_list = [ "apple", "blood", "rose", "strawberry", "cherry" ]
green_words_list = [ "grass", "tree", "forest", "clover" ]
yellow_words_list = [ "banana", "star", "sun", "desert", "lemon", "gold" ]

It will be used to colorize lyrics using my audioleds.py script.

  • Try wikidata -- start with the linked query and refine. – Barry Carter Dec 21 '17 at 12:13
  • seems a good start, if you know how to select only concrete words this could be a valid answer... – eadmaster Dec 22 '17 at 12:55
  • There is probably a wikidata property for concreteness, but I'm too lazy to check. – Barry Carter Dec 22 '17 at 14:04

I think if you rephrase your question, google gives some good results. For example, those in marketing & branding study the association of colors with feelings.

The article above has sources linked for everything, and depending on how wide you want to go from the literal color of the object, this could be a good direction.

Another direction could be linguistic research studies, where people associate words with colors.

To find this page, I had to start with a basic web search, and then remove false hits (like "Excel"). Here was the final query, maybe it helps.


(so I removed results with "excel", "Synesthesia", "highlight", and "html". I think in your question, using the word "mapping" will bring up mostly technical results, such as excel)


I won't expect a ready-made data set for these specifications.

However, it is easy create such kind of dataset yourself. You need a big corpus of English text and a Corpus Query Processor (like cqpweb). A query for red things than looks like

[word="red" [pos="N.*"]

asking for bigrammes consisting of the word "red" and a noun. Than, do a frequency breakdown to get output like this one (data are taken from the Corpus of Late Modern English CLMET):

Your query “[word="red"] [pos="N.*"]” returned 1,803 matches in 198 different texts (in 40,340,760 words [333 texts]; frequency: 44.69 instances per million words)
Showing frequency breakdown of words in this query, at the query node; there are 656 different types and 1,803 tokens at this concordance position.
No.     Search result   No. of occurrences  Percent
1   red hair    47  2.61%
2   red face    40  2.22%
3   red clay    35  1.94%
4   red deer    28  1.55%
5   red weed    27  1.5%
6   red brick   27  1.5%
7   red colour  26  1.44%
8   red glow    26  1.44%
9   red eyes    24  1.33%

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.